Do rats feel empathy?

12 05 2012

We read this controversial paper in Thursday’s journal club: Having heard criticisms of this work I was pleasantly surprised. What impressed me most was that rats would not just work to release a trapped rat into their own test arena (which could be explained by motivations for social contact): they would work to release a rat into a neighbouring arena. Cool. The authors also tried to deal with the criticism that that real reinforcer is the termination of distress signals: that rats work, not to release a captor, but to stop them making horrible distress calls. (After all, if I’m on a flight on its descent and the babies on board start crying, my desire for them to shut up is not driven by empathy….). However, we agreed that their attempts here were rather lame, and that the fact that the authors didn’t mention stress odours (let alone try and deal with them) was a major flaw ( As for the alleged chocolate-sharing? Meh. And to me the strangest gap was not comparing the non-releasers with the releasers: are the non-releasing rats…. not very bright? Autistic/uncaring? Did they have bad relationships with their cagemates (rat schadenfreude?!)? Or were their trapped cagemates quite content and not ‘asking’ for release? Overall, we felt these data were consistent with empathy, but not strong proof.

But everyone’s a critic right? And sniping is easy. How would WE go about testing this hypothesis? This was the fun part and here is what we came up with:

If rats really feel empathy….

– releasers’ behaviour should be driven by emotion; for example, indwelling heartrate monitors should show that a releaser’s stress response mirrors that of the captor; and captors who are highly stressed (perhaps because they have been preconditioned to associate the tube with shock or some other aversive stimulus) should be released faster than captors who are relaxed (e.g. because they have been preconditioned to associate the tube with being fed chocolate);

– former captors, especially those who had had a horrible time in their tubes, should more avid releasers of other rats than naive rats, or former captors who had had a nice, chocolate-related time in their tubes (this came from Dana….. and I love it). We imagined a 2 x 2 design using relaxed versus stressed captors, and releasers who had formerly been either relaxed or stressed captors: if empathy is the driver, it’s easy to predict which combination should lead to the most diligent releasing, and which should lead to little or none…

– cagemates who get on well should be more willing to release each other than cagemates who don’t (even into a different arena, so it’s definitely not motivated by social reinstatement); ‘getting on well’ could be quantified via in social interaction in the homecage, and better, assessing how hard cagemates will work to access each other when isolated (as in Anne Lene Hovland’s study of vixens:;

– releasing behaviour should be enhanced by oxytocin (Elena Choleris, who joined us from Psych., was keen on this:;

– other explanations need to be ruled out, e.g. the termination of stress odour release (one control might be to see if removing a source of conspecific stress odours is indeed a reinforcer), and ‘boredom’ (do groups of rats in enriched arenas still release captors?).